Fig. 1: Overview of scGCO for SV gene identification.
From: Identification of spatially variable genes with graph cuts
a A gene’s spatial expression pattern. Each dot represents a cell and is placed according to its spatial coordinate. b Representing a gene’s spatial expression with hidden Markov random field (HMRF). c Optimizing HMRF using graph cuts algorithm with different smooth factors and identifying the best graph cuts result that maximizes a score based on the signal-to-noise ratio. d P-value for each gene was evaluated using the best segmentation under the complete spatial randomness (CSR) framework. Benjamini–Hochberg (BH) correction was utilized to identify spatially variable (SV) genes at genome-scale. Cells are represented with Voronoi diagrams. Thicker lines highlight the segmentation boundaries identified by graph cuts.
